The trees can be fixed with re-prompting but my objective was to try and get it done in one shot.
Here is a Voxel Pagoda made with Qwen 3.8 Flash Next using laurent_zw 's quant with a Vulkan backend on Strix Halo avg 41 tok/s. Speed burst up to 47 tok/s using ROCmFPX build.
Testing Qwen 3.8 Flash running on 256gb m3 ultra.
Told it to make Halo 3 slayer with bots, make the map Guardian. Gave it reference images for the battle rifle and map. This took 2 hours and sat at about 120gb memory usage
China's new AI model shocks the internet
Qwen 3.8-Max-0902 hits 2.4 trillion parameters, built for coding and AI agents.
Qwen 3.8 Max vs Gemini 3.8 Flash vs Muse Spark 1.3.
Two identical visual coding tests. One response each. No tools, web, or repairs. Muse Spark won with 63.9/100. Three blank panels were retried, labeled, and not rescored.
You have reached the end of the archive
All of qwen38